1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
|
catdoc(1) catdoc(1)
NAME
catdoc - reads MS-Word file and puts its content as plain text on stan-
dard output
SYNOPSIS
catdoc [-vlu8btawxV] [-m number] [ -s charset] [ -d charset] [ -f out-
put-format] file
DESCRIPTION
catdoc behaves much like cat(1) but it reads MS-Word file and produces
human-readable text on standard output. Optionally it can use latex(1)
escape sequenses for characters which have special meaning for LaTeX.
It also makes some effort to recognize MS-Word tables, although it
never tries to write correct headers for LaTeX tabular environment.
Additional output formats, such is HTML can be easily defined.
catdoc doesn't attempt to extract formatting information other than
tables from MS-Word document, so different output modes means mainly
that different charachers should be escaped and different ways used to
represent characters, missing from output charset. See CHARACTER SUB-
STITUTION below
catdoc uses internal unicode(7) representation of text, so it is able
to convert texts when charset in source document doesn't match charset
on target system. See CHARACTER SETS below.
If no file names supplied, catdoc processes its standard input unless
it is terminal. It is unlikely that somebody could type Word document
from keyboard, so if catdoc invoked without arguments and stdin is not
redirected, it prints brief usage message and exits. Processing of
standard input (even among other files) can be forced using dash '-' as
file name.
By default, catdoc wraps lines which are more than 72 chars long and
separates paragraphs by blank lines. This behavoir can be turned of by
-w switch. In wide mode catdoc prints each paragraph as one long line,
suitable for import into word processors which perform word wrapping
theirselves.
OPTIONS
-a - shortcut for -f ascii. Produces ASCII text as output. Sepa-
rates table columns with TAB
-b - process broken MS-Word file. Normally, catdoc checks if first
8 bytes of file is Microsoft OLE signature. If so, it processes
file, otherwise it just copies it to stdin. It is intended to
use catdoc as filter for viewing all files with .doc extension.
-dcharset
- specifies destination charset name. Charset file has format
described in CHARACTER SETS below and should have .txt exten-
sion and reside in catdoc library directory ( ${exec_pre-
fix}/lib/catdoc). By default, current locale charset is used if
langinfo support compiled in.
-fformat
- specifies output format as described in CHARACTER SUBSTITU-
TION below. catdoc comes with two output formats - ascii and
tex. You can add your own if you wish.
-l Causes catdoc to list names of available charsets to the stdout
and exit successfully.
-mnumber
Specifies right margin for text (default 72). -m 0 is equiva-
lent to -w
-scharset
Specifies source charset. (one used in Word document), if Word
document doesn't contain UTF-16 text.
-t - shortcut for -f tex
converts all printable chars, which have special meaning for
LaTeX(1) into appropriate control sequences. Separates table
columns by &.
-u - declares that Word document contain UNICODE (UTF-16)
represntation of text (as some Word-97 documents). If catdoc
fails to correct Word document with default charset, try
this option.
-8 - declares is Word document is 8 bit. Just in case that catdoc
recognizes file format incorrectly.
-w disables word wrapping. By default catdoc output is splitted
into lines not longer than 72 (or number, specified by -m
option) characters and paragraphs are separated by blank
line. With this option each paragraph is one long line.
-x causes catdoc to output unknown UNICODE characher as \xNNNN,
instead of question marks.
-v causes catdoc to print some useless information about word doc-
ument structure to stdout before actual start of text.
-V outputs catdoc version
CHARACTER SETS
When processing MS-Word file catdoc uses information about two charac-
ter sets, typically different
- input and output. They are stored in plain text files in catdoc
data directory. Character set files should contain two whitespace-sepa-
rated hexadecimal numbers - 8-bit code in character set and 16-bit uni-
code code. Anything from hash mark to end of line is ignored, as well
as blank lines.
catdoc distribution includes some of these character sets. Additional
character set definitions, directly usable by catdoc can be obtained
from ftp.unicode.org. Charset files have .txt suffix, which shouldn't
be specified in command-line or configuration files.
Note that catdoc is distributed with Cyrillic charsets as default. If
you are not Russian, you probably don't want it, an should reconfigure
catdoc at compile time or in runtime configuration file.
When dealing with documents with charsets other than default, remember
that Microsoft never uses ISO charsets. While letters in, say cp1252
are at the same position as in ISO-8859-1, some punctuation signs would
be lost, if you specify ISO-8859-1 as input charset. If you use cp1252,
catdoc would deal with those signs as described in CHARACTER SUBSTITU-
TION below.
CHARACTER SUBSTITUTION
catdoc converts MS-Word file into following internal unicode represen-
tation:
1. Paragraphs are separated by ASCII Line Feed symbol (0x000A)
2. Table cells within row are separated by ASCII Field Separator symbol
(0x001C)
3. Table rows are separated by ASCII Record Separator (0x001E)
4. All printable characters, including whitespace are represented with
their
respective UNICODE codes.
This UNICODE representation is subsequentely converted into 8-bit text
in target character set using following four-step algorithm:
1. List of special characters is searched for given unicode character.
If found, then appropriate multi-character sequence is output
instead of character.
2. If there is an equivalent in target character set, it is output.
3. Otherwise, replacement list is searched and, if there is multi-char-
acter
substitution for this UNICODE char, it is output.
4. If all above fails, "Unknown char" symbol (question mark) is output.
Lists of special characters and list of substitution are character set-
independent, becouse special chars should be escaped regardless of
their existense in target character set (usially, they are parts of
US-ASCII, and therefore exist in any character set) and replacement
list is searched only for those characters, which are not found in tar-
get character set.
These lists are stored in catdoc data directory in files with prefix of
format name. These files have following format:
Each line can be either comment (starting with hash mark) or contain
hexadecimal UNICODE value, separated by whitespace from string, which
would be substituted instead of it. If string contain no whitespace it
can be used as is, otherwise it should be enclosed in single or double
quotes. Usial backslash sequences like '\n','\t' can be used in these
string.
RUNTIME CONFIGURATION
Upon startup catdoc reads its system-wide configuration file /etc/cat-
docrc and then user-specific configuration file ${HOME}/.catdocrc.
These files can contain following directives:
source_charset = charset-name
Sets default source charset, which would be used if no -s
option specified. Consult configuration of nearby windows work-
station to find one you need.
target_charset = charset-name
Sets default output charset. You probably know, which one you
use.
charset_path = directory-list
colon-separated list of directories, which are searched for
charset files. This allows you to install additional charsets
in your home directory.
map_path = directory-list
colon-separated list of directories, which are searched for
special character map and replacement map.
format = format name
Output format which would be used by default. catdoc comes
with two formats - ascii and tex but nothing prevents you from
writing your own format (set two map files - special character
map and replacement map).
unknown_char = character specification
sets characher to output instead of unknown unicode character
(default '?') Character specification can have one of two form
- character enclosed in single quotes or hexadecimal code.
use_locale =(yes|no)
Enables or disables automatic selection of output charset
(default yes),
based on system locale settings (if enabled at compile time).
If automatic detection is enabled, than output charset settings
in the configuration files (but not in the command line) are
ignored, and current system locale charset is used instead.
There are no automatic choice of input charset, based of locale
language, because most modern Word files (since Word 97) are
Unicode anyway
BUGS
Doesn't handle fast-saves properly. Prints footnotes as separate para-
graphs at the end of file, instead of producing correct LaTeX commands.
Cannot distinguish between empty table cell and end of table row.
SEE ALSO
xls2csv(1), cat(1), strings(1), utf(4), unicode(7)
AUTHOR
V.B.Wagner <vitus@ice.ru>
MS-Word reader Version 0.93.4 catdoc(1)
|
